Sliding mode control for singularly perturbed systems

The design of sliding mode control for singularly perturbed systems is investigated in this thesis. A number of methods of sliding mode control for singularly perturbed system are presented. A single sliding mode controller has been proposed to control the full order singularly perturbed system. The design method used for this controller is based on the reduced mathematical model of both the slow and the fast subsystems. The design objective is to achieve the desired performance with feedback based only on the slow states.; An accurate decoupling method, known as the diagonalization method, is used to obtain the reduced order fast and slow subsystems. We design the single sliding mode controller based on the slow subsystem. We discuss and prove convergence and stability for the proposed controller, and present some numerical results. A maximum limit for the perturbation parameter epsilon that guarantee the stability of the proposed controller is shown. The sliding mode Hinfinity frequency shaped control method is used for the proposed controller to design a frequency shaped sliding surface. By using this technique we can suppress the high frequency modes of the fast subsystem. The fuzzy sliding mode control algorithm is applied to the proposed controller. The fuzzy logic acts as a supervisory system to connect two different sliding mode controllers together, which allows us to obtain a nonlinear sliding surface.; Examples that demonstrate the usefulness of these procedures are presented. The results indicate that the proposed controller has similar or better performance compared to other control strategies.